The Configuration File

Mr.Docs uses a configuration file to control how the documentation is generated. The file is used to specify options such as the generator to use, additional compilation options, and filters.

Here’s an example of a configuration file:

source-root: ../include
multipage: false
generator: adoc

The Usage page provides a detailed explanation of what to combine options from the configuration file and the command line.

The Reference section provides a detailed explanation of the options available.

More information about the generators can be found in the Generators page.

YAML Schema

To get linting and autocompletion in the config file, a schema for the config can be specified. The schema for mrdocs.yml is provided here.

The schema for mrdocs.yml is available from https://www.schemastore.org/json/, which is automatically detected and used by most editors.

To manually set the schema in an editor, the following can be used:

In JetBrains IDEs, # $schema: <url> can be used to bind a schema to a file.
In editors with plugins based on the YAML language server, # yaml-language-server: $schema=<url> can be used.

The following shows an example of a file specifying an inline-schema that’s compatible with JetBrains IDEs and editors using the YAML language server.

# $schema: https://mrdocs.com/docs/mrdocs/develop/_attachments/mrdocs.schema.json
# yaml-language-server: $schema=https://mrdocs.com/docs/mrdocs/develop/_attachments/mrdocs.schema.json

source-root: ../include
multipage: false
generator: adoc

Build Options

A number of options can be used to specify with which compile options Mr.Docs should be run.

source-root: ..
compilation-database: ../CMakeLists.txt
cmake: '-D MRDOCS_BUILD=ON'
defines: 'MRDOCS_BUILD'

The compile options primarily come from the compilation-database file. When this file is generated from a CMakeLists.txt script, the cmake option can be used to specify additional options to pass to CMake.

Additionally, the defines option can be used to specify preprocessor definitions that should be used when generating the documentation. These definitions are included in all targets of the compilation database.

Filters

Not all symbols in a project may be relevant to the documentation. Mr.Docs provides a way to filter out symbols based on their location and names.

File Filters

Symbols can also be filtered based on the files where they are declared. This can be useful for excluding files that exclusively contain implementation details or test code.

The input option is a list of directories to include and a list of file patterns to include. Only symbols in files from these directories will be extracted.

input:
  - ../include

The default value for input is "<source-root>/.", so only symbols defined in your project source directory will be extracted.

The file-patterns options allows you to specify a list of glob file patterns to include. Only symbols in files whose filename match these patterns will be extracted.

file-patterns:
  - '*.hpp'
  - '*.h'

The default value for file-patterns include a variety of common C++ header file extensions.

The exclude option is a list of subdirectories in input we want to exclude. Meanwhile, exclude-patterns can use glob patterns to exclude files from the extraction process.

input:
  - ../include
exclude:
  - ../include/detail
exclude-patterns:
  - ../include/*/detail/**
  - ../include/*/impl/**

The glob patterns support the following wildcards:

* matches all characters except delimiters /
** matches all characters
? matches any single character.
[<chars>] matches one character in the bracket.
[<char>-<char>] matches one character in the bracket range.
[^<chars>] or [!<chars>] matches one character not in the bracket.
{<glob>,…} matches one of the globs in the list.
\ escapes the next character so it is treated as a literal.

Symbol Filters

Symbols can also be filtered based on their qualified names. You can use glob patterns to specify which symbols to include or exclude.

include-symbols:
  - 'my_library::public_namespace::*'

By default, all symbols are included by include-symbols. The syntax for symbol glob patterns is the same as for file patterns, with the exception that the delimiter :: is used instead of /. So you can match all symbols in a namespace and its subnamespaces with my_library::public_namespace::**.

The option exclude-symbols can be used to exclude symbols from the extraction process.

include-symbols:
  - 'my_library::**'
exclude-symbols:
  - 'my_library::private_namespace::**'

// excluded by `include-symbols`
void f0();

// included because it matches the prefix of 'my_library::**' in `include-symbols`
namespace my_library
{
    // included because it matches the pattern 'my_library::**' in `include-symbols`
    void f1();

    namespace private_namespace
    {
        // excluded by the pattern 'my_library::private_namespace::**' in `exclude-symbols`
        void f2();
    }
}

Private Symbols

The implementation-defined and see-below options can be used to designate symbols as implementation details or "see below" in the documentation.

include-symbols:
  - 'my_library::**'
implementation-defined:
  - 'my_library::detail::**'
see-below:
  - 'my_library::see_below::**'

namespace my_library
{
    namespace detail
    {
        // There's no documentation page for A
        // Any reference to `A` is rendered as `/* implementation detail */`
        class A {};
    }

    namespace see_below
    {
        // The documentation page for B is marked as "see below" and its members are not extracted.
        class B {
            // There's no documentation page for iterator
            class iterator;
        };
    }

    // This function is documented, but the return type is rendered as `/* implementation detail */`
    detail::A
    foo();
}

Whitelisting Rules

The rules for whitelisting symbols (include-symbols, implementation-defined, and see-below) are less strict than the rules for blacklisting symbols (exclude-symbols). A symbol is considered whitelisted if it matches any of the following conditions:

The symbol strictly matches one of the patterns.
- For instance, the patterns std::vector and std::* both match std::vector strictly.
The symbol is a parent namespace of an included symbol.
- For instance, the pattern std::filesystem::* also includes std and std::filesystem.
The parent symbol is also included.
- For instance, the pattern std::* also matches std::vector::iterator because std::vector::iterator is a member of std::vector, which is matches the pattern.
The symbol is a child of a literal pattern representing a namespace.
- For instance, the literal pattern std matches std::filesystem::path::iterator because std is a literal pattern matching a namespace. In other words, these literal patterns represent the namespace and its subnamespaces as if the pattern were std::**.

For exclusion rules, the symbol must strictly match the pattern to be excluded. If a scope is escaped by a pattern, all symbols in that scope are also excluded.

Reference

The following options can be defined both in the configuration file and on the command line, where the command line options always take precedence.

Whenever applicable, the command line options attempt to mirror (i) non-abbreviated (ii) kebab-case variants of the equivalent commands in Doxygen. For instance, the Doxygen EXTRACT_ANON_NSPACES option is mirrored by the extract-anonymous-namespaces command line option.

Command Line Options

Options that can only be provided via the command line

The following options can be used to control the general behavior of Mr.Docs and can only be provided via the command line. These include options to specify inputs and the configuration file, which cannot be set on the configuration file itself.

Name Description Default

cmd-line-inputs
(list of paths)
(Command line only) Configuration or compilation database files []

config
(file path)
(Required, Command line only) Mr.Docs configuration file "<cwd>/mrdocs.yml"

Name	Description	Default
`cmd-line-inputs` (list of paths) (Command line only)	Configuration or compilation database files	[]
`config` (file path) (Required, Command line only)	Mr.Docs configuration file	"<cwd>/mrdocs.yml"

cmd-line-inputs

Configuration or compilation database files

The inputs are configuration files or compilation database files that used to generate the documentation. When the input ends with mrdocs.yml, it is interpreted as a configuration file, the file is read and the options are used to generate the documentation as if it was provided to the config option. When the input ends with compilation_database.json or CMakeLists.txt, it is interpreted as a compilation database file, the file is read and the compiler flags are used to generate the documentation as if it was provided to the compilation-database option.

Type: list of paths
Command line only
Default value: []
This command is a command line sink. Any command line argument that is not recognized by the parser will be passed to this command.

config

Mr.Docs configuration file

The configuration file is a YAML file that contains the options used to generate the documentation. The configuration file is read and the options are used to generate the documentation. The configuration file can be used to specify the source code, the output directory, the compilation database, the generator, and the filters.

Type: file path
Required
Command line only
Default value: "<cwd>/mrdocs.yml"

Paths

Paths to the source code and output directories

Name Description Default

source-root
(directory path) Path to the root directory of the source code "<config-dir>"

output
(path) Directory or file for generating output "<config-dir>/reference-output"

compilation-database
(file path) Path to the compilation database

Name	Description	Default
`source-root` (directory path)	Path to the root directory of the source code	"<config-dir>"
`output` (path)	Directory or file for generating output	"<config-dir>/reference-output"
`compilation-database` (file path)	Path to the compilation database

source-root

Path to the root directory of the source code

Path to the root directory of the source code. This path is used as a default for input files and a base for relative paths formed from absolute paths. This should typically be the root directory of the git project, as relative paths formed from it can be used to create links to these source files in the repository. Templates use the base-url option to create links to the source code.

Type: directory path
Default value: "<config-dir>"

output

Directory or file for generating output

Multipage generators expect a directory. Single page generators expect a file or a directory where the file will be created. If the directory does not exist, it will be created.

Type: path
Default value: "<config-dir>/reference-output"

compilation-database

Path to the compilation database

Path to the compilation database or a build script to generate it. The compilation database is a JSON file that contains the compiler commands used to build the source code. The compilation database is used to extract the compiler flags and the source files used to build the source code and extract symbols. This option also accepts the path to a build script such as CMakeLists.txt to be used to generate the compilation database. In this case, Mr.Docs will look for CMake in PATH or in CMAKE_ROOT and run the script to generate the compilation database file.

Type: file path
Default value:

Filters

Filters to include or exclude files and symbols from the documentation

Name Description Default

input
(list of paths) Input directories to extract symbols from ["<source-root>/."]

recursive
(boolean) Recursively include files from "input" paths true

file-patterns
(list of path-globs) File patterns to include ["*.hpp", "*.h", "*.hh", "*.ipp", "*.inc", "*.cpp", "*.cc", "*.cxx", "*.c", "*.hxx"]

exclude
(list of paths) Input directories to exclude []

exclude-patterns
(list of path-globs) File patterns to exclude []

include-symbols
(list of symbol-globs) Symbol patterns to include []

exclude-symbols
(list of symbol-globs) Symbol patterns to exclude []

see-below
(list of symbol-globs) Exposition only symbols rendered as "see-below". []

implementation-defined
(list of symbol-globs) Symbols rendered as "implementation-defined" []

Name	Description	Default
`input` (list of paths)	Input directories to extract symbols from	["<source-root>/."]
`recursive` (boolean)	Recursively include files from "input" paths	true
`file-patterns` (list of path-globs)	File patterns to include	[".hpp", ".h", ".hh", ".ipp", ".inc", ".cpp", ".cc", ".cxx", ".c", ".hxx"]
`exclude` (list of paths)	Input directories to exclude	[]
`exclude-patterns` (list of path-globs)	File patterns to exclude	[]
`include-symbols` (list of symbol-globs)	Symbol patterns to include	[]
`exclude-symbols` (list of symbol-globs)	Symbol patterns to exclude	[]
`see-below` (list of symbol-globs)	Exposition only symbols rendered as "see-below".	[]
`implementation-defined` (list of symbol-globs)	Symbols rendered as "implementation-defined"	[]

input

Input directories to extract symbols from

Input directories to extract. Only symbols defined in files in these directories are extracted. The paths are relative to the mrdocs configuration file.

Type: list of paths
Default value: ["<source-root>/."]

Example

include/libavcodec/avcodec.h

/** Allocate a new packet on the heap. */
int av_packet_alloc(void);

/** Release a packet allocated with `av_packet_alloc`. */
void av_packet_free(void);

include/libavformat/avformat.h

// Filtered out: `input: include/libavcodec` scopes extraction to that
// directory, so the format-library functions never enter the corpus.
// The directory matters because both libraries put their declarations
// at global scope under the same `av_*` prefix, so a name-based filter
// could not distinguish them.
int av_open_input_file(void);
void av_close_input_file(void);

input.cpp

#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>

mrdocs.yml

input:
  - include/libavcodec

Preview·input

av_packet_alloc

Allocate a new packet on the heap.

Synopsis

Declared in <libavcodec/avcodec.h>

int
av_packet_alloc();

av_packet_free

Release a packet allocated with av_packet_alloc.

Synopsis

Declared in <libavcodec/avcodec.h>

void
av_packet_free();

recursive

Recursively include files from "input" paths

Recursively include files. When set to true, Mr.Docs includes files in subdirectories of the input directories. When set to false, Mr.Docs includes only the files in the input directories.

Type: boolean
Default value: true

Example

include/cli/detail/tokens.hpp

namespace cli::detail {
// Filtered out: lives in a subdirectory that `recursive: false`
// skips, so `cli::detail::tokenize` never enters the corpus.
void tokenize(char const* arg);
}

include/cli/parse.hpp

namespace cli {
/** Parses command-line arguments into an option set. */
void parse(int argc, char** argv);
}

recursive.cpp

#include <cli/parse.hpp>
#include <cli/detail/tokens.hpp>

mrdocs.yml

input:
  - include/cli
recursive: false

Preview·recursive

cli::parse

Parses command‐line arguments into an option set.

Synopsis

Declared in <cli/parse.hpp>

void
parse(
    int argc,
    char** argv);

file-patterns

File patterns to include

File patterns to include. Only the files that match these patterns are extracted. The patterns are relative to the input directories.

Type: list of path-globs
Default value: ["*.hpp", "*.h", "*.hh", "*.ipp", "*.inc", "*.cpp", "*.cc", "*.cxx", "*.c", "*.hxx"]

exclude

Input directories to exclude

Symbols defined in files in these directories are not extracted even if they are in the list of include directories. When relative, the paths are relative to the directory of the mrdocs configuration file. For instance, "include/experimental" will exclude all files in the directory <config-dir>/include/experimental.

Type: list of paths
Default value: []

Example

include/httpd/server.hpp

namespace httpd {

/** Start the HTTP server on `port`. */
void start(int port);

/** Stop the running server. */
void stop();

}

include/zlib/zlib.h


// Vendored copy of zlib's public API. Documented upstream at
// https://zlib.net/, not here. `exclude: include/zlib` keeps the
// vendored library out of the rendered docs so it does not appear
// alongside the httpd API.

/* Compress `src` into `dst` using DEFLATE. */
int compress2(unsigned char* dst, unsigned long* dstLen,
              unsigned char const* src, unsigned long srcLen,
              int level);

/* Decompress a DEFLATE-compressed buffer. */
int uncompress(unsigned char* dst, unsigned long* dstLen,
               unsigned char const* src, unsigned long srcLen);

exclude.cpp

#include <httpd/server.hpp>
#include <zlib/zlib.h>

mrdocs.yml

exclude:
  - include/zlib

Preview·exclude

httpd::start

Start the HTTP server on port.

Synopsis

Declared in <httpd/server.hpp>

void
start(int port);

httpd::stop

Stop the running server.

Synopsis

Declared in <httpd/server.hpp>

void
stop();

exclude-patterns

File patterns to exclude

File patterns to exclude. Files that match these patterns are not extracted even if they are in the list of include directories. The patterns are relative to the configuration file. A single * will match all files in the directory. Double ** will match all files in the directory and its subdirectories.

Type: list of path-globs
Default value: []

Example

include/payments/_generated/messages.hpp

namespace payments::gen {
struct AuthorizeRequest {};
struct AuthorizeResponse {};
}

include/payments/payments.hpp

namespace payments {

/** Authorize a payment of `amount` against `account`. */
bool authorize(char const* account, double amount);

}

include/storage/_generated/messages.hpp


// Generated by the build (protoc, gRPC, OpenAPI, …); do not edit by
// hand. Every component in this project keeps its generated headers
// next to the hand-written ones under a `_generated/` subdirectory,
// so `exclude-patterns: '**/_generated/**'` drops them all at once.

namespace storage::gen {
struct PutRequest {};
struct PutResponse {};
}

include/storage/storage.hpp

namespace storage {

/** Persist a value under `key`. */
void put(char const* key, char const* value);

}

exclude-patterns.cpp

#include <storage/storage.hpp>
#include <storage/_generated/messages.hpp>
#include <payments/payments.hpp>
#include <payments/_generated/messages.hpp>

mrdocs.yml

exclude-patterns:
  - '**/_generated/**'

Preview·exclude-patterns

payments::authorize

Authorize a payment of amount against account.

Synopsis

Declared in <payments/payments.hpp>

bool
authorize(
    char const* account,
    double amount);

storage::put

Persist a value under key.

Synopsis

Declared in <storage/storage.hpp>

void
put(
    char const* key,
    char const* value);

include-symbols

Symbol patterns to include

If any patterns are defined here, only symbols that match one of these patterns are extracted. The patterns are applied to the fully qualified name of the symbol without any leading "::". A single "*" will match all symbols in the namespace. Double "**" will match all symbols in the namespace and its subnamespaces. The patterns also support "?" for any chars, "[<chars>]" for charsets, "[^<chars>]" for inverted charsets, and "{<glob>,...}" for alternatives.

Type: list of symbol-globs
Default value: []

Example

Input

namespace lib {
namespace detail {
/// An internal helper that should not appear in the docs.
int helper();
}

/// The library's only public entry point.
int run();
}

mrdocs.yml

include-symbols:
  - 'lib::run'

Preview·include-symbols

lib::run

The library's only public entry point.

Synopsis

Declared in <include‐symbols.cpp>

int
run();

exclude-symbols

Symbol patterns to exclude

A symbol that matches one of these patterns is not extracted even if whitelisted by "include-symbols". See the documentation for "include-symbols" for the pattern syntax.

Type: list of symbol-globs
Default value: []

Example

include/jzon/extra/utilities.hpp

namespace jzon::extra {

// Filtered out: `exclude-symbols: 'jzon::extra::**'` drops these
// symbols by qualified name, so the `jzon::extra` namespace and its
// members never enter the rendered docs.
void enable_trace();
void disable_trace();

}

include/jzon/parser.hpp

namespace jzon {

/** Validate that the document is well-formed JSON.

    Parses `source` without constructing a value tree, reporting only
    whether the input conforms to RFC 8259. Comments and trailing
    commas are rejected.

    @param source A null-terminated UTF-8 JSON document.
    @return `true` if the document is syntactically valid JSON, `false`
            if any syntactic error is encountered.
*/
bool validate(char const* source);

}

exclude-symbols.cpp

#include <jzon/parser.hpp>
#include <jzon/extra/utilities.hpp>

mrdocs.yml

exclude-symbols:
  - 'jzon::extra::**'

Preview·exclude-symbols

jzon::validate

Validate that the document is well‐formed JSON.

Synopsis

Declared in <jzon/parser.hpp>

bool
validate(char const* source);

Description

Parses source without constructing a value tree, reporting only whether the input conforms to RFC 8259. Comments and trailing commas are rejected.

Return Value

true if the document is syntactically valid JSON, false if any syntactic error is encountered.

Parameters

Name

Description

source

A null‐terminated UTF‐8 JSON document.

see-below

Exposition only symbols rendered as "see-below".

Symbols that match one of these filters are tagged as "see-below" in the documentation, and so do symbols in scopes tagged as "see-below". This option is used to remove details about symbols that are considered part of the private API of the project but the user might need to interact with. In the documentation page for this symbol, the symbol is exposition only: the synopsis of the implementation is rendered as "see-below" and members of scopes (such as a namespace or record) are not listed. The rest of the documentation is rendered as usual to explain the symbol. See the documentation for "include-symbol" for the pattern syntax.

Type: list of symbol-globs
Default value: []

Example

include/netz/buffer_view.hpp

namespace netz {

/** A non-owning view of a contiguous byte range.

    Used as the parameter and return type for the framing helpers in
    `netz::stream` and as the buffer surface presented to async read
    handlers. Stores a pointer and a length, and does not extend the
    lifetime of the storage it refers to.
*/
class buffer_view
{
public:
    unsigned char const* data() const noexcept;
    unsigned long size() const noexcept;
};

}

include/netz/stream.hpp

namespace netz {

/** Locate the first complete packet at the start of the buffer.

    Scans `input` for a frame boundary. The returned view aliases the
    same memory as `input`, so the storage behind `input` must outlive
    the returned view.

    @param input Bytes received from the stream, up to the caller's
                 current fill mark.
    @return A view of the bytes belonging to the first packet, or an
            empty view if no complete packet is present yet.
*/
buffer_view first_packet(buffer_view input);

}

see-below.cpp

#include <netz/stream.hpp>

mrdocs.yml

see-below:
  - 'netz::buffer_view'

Preview·see-below

netz::buffer_view

A non‐owning view of a contiguous byte range.

Synopsis

Declared in <netz/buffer_view.hpp>

class buffer_view { /* see-below */ };

Description

Used as the parameter and return type for the framing helpers in netz::stream and as the buffer surface presented to async read handlers. Stores a pointer and a length, and does not extend the lifetime of the storage it refers to.

netz::first_packet

Locate the first complete packet at the start of the buffer.

Synopsis

Declared in <netz/stream.hpp>

buffer_view
first_packet(buffer_view input);

Description

Scans input for a frame boundary. The returned view aliases the same memory as input, so the storage behind input must outlive the returned view.

Return Value

A view of the bytes belonging to the first packet, or an empty view if no complete packet is present yet.

Parameters

Name

Description

input

Bytes received from the stream, up to the caller's current fill mark.

implementation-defined

Symbols rendered as "implementation-defined"

Symbols that match one of these filters are tagged as "implementation-defined" in the documentation, and so do symbols in scopes tagged as "implementation-defined". This option is used to exclude symbols from the documentation that are considered part of the private API of the project. An "implementation-defined" symbol has no documentation page in the output. If any other symbol refers to it, the reference is rendered as "implementation-defined". See the documentation for "include-symbol" for the pattern syntax.

Type: list of symbol-globs
Default value: []

Example

include/logr/detail/scope_token.hpp

namespace logr::detail {
struct scope_token { ~scope_token(); };
}

include/logr/log.hpp

namespace logr {

/** Attach a key/value pair to every log line in the current scope.

    The context stays attached until the returned token is destroyed.
    Hold it in `auto`; nested contexts stack and unwind in reverse order.

    @par Example
    @code
    void handle_request(request const& req) {
        auto ctx = logr::scoped_context("request_id", req.id);
        do_work(req);
    }
    @endcode

    @param key   The context name surfaced on each log line.
    @param value The context value surfaced on each log line.
    @return An opaque RAII token whose lifetime governs how long the pair
            stays attached. The type is implementation-defined; store it
            in `auto`.
*/
detail::scope_token scoped_context(char const* key, char const* value);

}

implementation-defined.cpp

#include <logr/log.hpp>

mrdocs.yml

implementation-defined:
  - 'logr::detail::**'

Preview·implementation-defined

logr::scoped_context

Attach a key/value pair to every log line in the current scope.

Synopsis

Declared in <logr/log.hpp>

/* implementation-defined */
scoped_context(
    char const* key,
    char const* value);

Description

The context stays attached until the returned token is destroyed. Hold it in auto; nested contexts stack and unwind in reverse order.

Example

void handle_request(request const& req) {
    auto ctx = logr::scoped_context("request_id", req.id);
    do_work(req);
}

Return Value

An opaque RAII token whose lifetime governs how long the pair stays attached. The type is implementation‐defined; store it in auto.

Parameters

Name

Description

key

The context name surfaced on each log line.

value

The context value surfaced on each log line.

Semantic Parsing

Options to control how the source code is parsed

Mr.Docs uses libclang to parse the source code and extract symbols. The following options control how the source code is parsed.

Name Description Default

missing-include-prefixes
(list of strings) Include path prefixes allowed to be missing []

missing-include-shims
(map) Shims for forgiven missing include files {}

Name	Description	Default
`missing-include-prefixes` (list of strings)	Include path prefixes allowed to be missing	[]
`missing-include-shims` (map)	Shims for forgiven missing include files	{}

missing-include-prefixes

Include path prefixes allowed to be missing

Specifies path prefixes for include files that, if missing, will not cause documentation generation to fail. Missing files with these prefixes are served as empty files from an in-memory file system, allowing processing to continue. For example, use "llvm/" to forgive all includes from LLVM. If any such path is specified, MrDocs will attempt to synthesize missing included types. Only simple sets of non-conflicting inferred types can be synthesized. For more complex types or for better control, provide a shim using the "missing-include-shims" option.

Type: list of strings
Default value: []

Example

mrdocs.yml

missing-include-prefixes:
  - "llvm/"
  - "boost/"

missing-include-shims

Shims for forgiven missing include files

Specifies a map of include file paths to shim contents. If a missing include file matches a forgiven prefix, MrDocs will use the shim content from this map as the file contents. If no shim is provided for a forgiven file, an empty file is used by default.

Type: map
Default value: {}

Example

mrdocs.yml

missing-include-shims:
  "llvm/ADT/StringRef.h": |
    namespace llvm { class StringRef; }
  "boost/version.hpp": |
    #define BOOST_VERSION 108400

Comment Parsing

Options to control how comments are parsed

Mr.Docs extracts metadata from the comments in the source code. The following options control how comments are parsed.

Name Description Default

auto-brief
(boolean) Use the first line of the comment as the brief true

auto-relates
(boolean) Automatically find non-member functions true

auto-function-metadata
(boolean) Automatically provide missing documentation for special functions and trivial metadata true

Name	Description	Default
`auto-brief` (boolean)	Use the first line of the comment as the brief	true
`auto-relates` (boolean)	Automatically find non-member functions	true
`auto-function-metadata` (boolean)	Automatically provide missing documentation for special functions and trivial metadata	true

auto-brief

Use the first line of the comment as the brief

When set to true, Mr.Docs uses the first line (until the first dot, question mark, or exclamation mark) of the comment as the brief of the symbol. When set to false, a explicit @brief command is required.

Type: boolean
Default value: true

Example

Input

/** Truncate a string to the given length.

    The first sentence of this comment becomes the
    brief; the rest is the description.
*/
void truncate(char* s, int n);

mrdocs.yml

auto-brief: true

Preview·auto-brief

truncate

Truncate a string to the given length.

Synopsis

Declared in <auto‐brief.cpp>

void
truncate(
    char* s,
    int n);

Description

The first sentence of this comment becomes the brief; the rest is the description.

auto-relates

Automatically find non-member functions

When set to true, Mr.Docs automatically finds non-member functions that are related to the current class.

Type: boolean
Default value: true

Example

Input

/** A two-dimensional point.
*/
struct point
{
    double x;
    double y;
};

/** Print a point to stdout.

    The free function takes `point` as its only
    parameter, so it gets attached to `point` as a
    related function.
*/
void print(point p);

mrdocs.yml

auto-relates: true

Preview·auto-relates

point

A two‐dimensional point.

Synopsis

Declared in <auto‐relates.cpp>

struct point;

Data Members

Name

x

y

Non-Member Functions

Name

Description

print

Print a point to stdout.

point::x

Synopsis

Declared in <auto‐relates.cpp>

double x;

point::y

Synopsis

Declared in <auto‐relates.cpp>

double y;

print

Print a point to stdout.

Synopsis

Declared in <auto‐relates.cpp>

void
print(point p);

Description

The free function takes point as its only parameter, so it gets attached to point as a related function.

Parameters

Name

Description

A two‐dimensional point.

auto-function-metadata

Automatically provide missing documentation for special functions and trivial metadata

When set to true, Mr.Docs automatically provides documentation for special functions, such as constructors, destructors, and operators. It also provides documentation for missing documentation metadata, such as known types.

Type: boolean
Default value: true

Example

Input

/// A 2D vector.
struct vec2
{
    /// Construct from `x` and `y` components.
    vec2(double x, double y);

    vec2(vec2 const&);
    vec2(vec2&&) noexcept;
    ~vec2();

    vec2& operator=(vec2 const&);

    bool operator==(vec2 const&) const;

    /// Returns the Euclidean length of the vector.
    double magnitude() const;

    /// Translate this vector by `delta`.
    vec2 translated(vec2 const& delta) const;
};

mrdocs.yml

auto-function-metadata: true

Preview·auto-function-metadata

vec2

A 2D vector.

Synopsis

Declared in <auto‐function‐metadata.cpp>

struct vec2;

Member Functions

Name

Description

vec2 [constructor]

Construct from x and y components.

~vec2 [destructor]

Destructor

operator=

Copy assignment operator

magnitude

Returns the Euclidean length of the vector.

translated

Translate this vector by delta.

operator==

Equality operator

vec2::vec2

Construct from x and y components.

Synopses

Declared in <auto‐function‐metadata.cpp>

Copy constructor

vec2(vec2 const& other);

» more...

Move constructor

vec2(vec2&& other) noexcept;

» more...

Construct from x and y components.

vec2(
    double x,
    double y);

» more...

vec2::vec2

Copy constructor

Synopsis

Declared in <auto‐function‐metadata.cpp>

vec2(vec2 const& other);

Parameters

Name

Description

other

The object to copy construct from

vec2::vec2

Move constructor

Synopsis

Declared in <auto‐function‐metadata.cpp>

vec2(vec2&& other) noexcept;

Parameters

Name

Description

other

The object to move construct from

vec2::vec2

Construct from x and y components.

Synopsis

Declared in <auto‐function‐metadata.cpp>

vec2(
    double x,
    double y);

Parameters

Name

Description

The value to construct from

vec2::~vec2

Destructor

Synopsis

Declared in <auto‐function‐metadata.cpp>

~vec2();

vec2::operator=

Copy assignment operator

Synopsis

Declared in <auto‐function‐metadata.cpp>

vec2&
operator=(vec2 const& other);

Return Value

Reference to the current object

Parameters

Name

Description

other

The object to copy assign from

vec2::magnitude

Returns the Euclidean length of the vector.

Synopsis

Declared in <auto‐function‐metadata.cpp>

double
magnitude() const;

Return Value

the Euclidean length of the vector.

vec2::translated

Translate this vector by delta.

Synopsis

Declared in <auto‐function‐metadata.cpp>

vec2
translated(vec2 const& delta) const;

Return Value

A 2D vector.

Parameters

Name

Description

delta

A 2D vector.

vec2::operator==

Equality operator

Synopsis

Declared in <auto‐function‐metadata.cpp>

bool
operator==(vec2 const& rhs) const;

Return Value

true if the objects are equal, false otherwise

Parameters

Name

Description

rhs

The right operand

Metadata Extraction

Metadata and C++ semantic constructs to extract

MrDocs extracts metadata and C++ semantic constructs from the source code to create the documentation. Semantic constructs are patterns not directly represented in the source code AST but can be inferred from the corpus, such as SFINAE. The following options control the extraction of metadata and C++ semantic constructs.

Name Description Default

extract-all
(boolean) Extract all symbols true

extract-private
(boolean) Extraction policy for private class members false

extract-private-virtual
(boolean) Extraction policy for private virtual methods of a class false

extract-private-bases
(boolean) Extraction policy for private base classes false

extract-static
(boolean) Extraction policy for static members of a file false

extract-local-classes
(boolean) Extraction policy for records defined locally in source files true

extract-anonymous-namespaces
(boolean) Extraction policy for anonymous namespaces true

extract-empty-namespaces
(boolean) Extraction policy for empty namespaces false

inherit-base-members
(enum) Determine how derived classes inherit base members "copy-dependencies"

extract-implicit-specializations
(boolean) Implicit template specializations used as base classes are extracted as dependencies true

extract-friends
(boolean) Extraction policy for friend functions and classes true

sort-members
(boolean) Sort the members of a record true

sort-members-by
(enum) Determine how members of a record are sorted "name"

sort-namespace-members-by
(enum) Determine how members of a namespace are sorted "name"

sort-members-ctors-1st
(boolean) Sort constructors first true

sort-members-dtors-1st
(boolean) Sort destructors first true

sort-members-assignment-1st
(boolean) Sort assignment operators first true

sort-members-conversion-last
(boolean) Sort conversion operators last true

sort-members-relational-last
(boolean) Sort relational operators last true

Name	Description	Default
`extract-all` (boolean)	Extract all symbols	true
`extract-private` (boolean)	Extraction policy for private class members	false
`extract-private-virtual` (boolean)	Extraction policy for private virtual methods of a class	false
`extract-private-bases` (boolean)	Extraction policy for private base classes	false
`extract-static` (boolean)	Extraction policy for static members of a file	false
`extract-local-classes` (boolean)	Extraction policy for records defined locally in source files	true
`extract-anonymous-namespaces` (boolean)	Extraction policy for anonymous namespaces	true
`extract-empty-namespaces` (boolean)	Extraction policy for empty namespaces	false
`inherit-base-members` (enum)	Determine how derived classes inherit base members	"copy-dependencies"
`extract-implicit-specializations` (boolean)	Implicit template specializations used as base classes are extracted as dependencies	true
`extract-friends` (boolean)	Extraction policy for friend functions and classes	true
`sort-members` (boolean)	Sort the members of a record	true
`sort-members-by` (enum)	Determine how members of a record are sorted	"name"
`sort-namespace-members-by` (enum)	Determine how members of a namespace are sorted	"name"
`sort-members-ctors-1st` (boolean)	Sort constructors first	true
`sort-members-dtors-1st` (boolean)	Sort destructors first	true
`sort-members-assignment-1st` (boolean)	Sort assignment operators first	true
`sort-members-conversion-last` (boolean)	Sort conversion operators last	true
`sort-members-relational-last` (boolean)	Sort relational operators last	true

extract-all

Extract all symbols

When set to true, MrDocs extracts all symbols from the source code, even if no documentation is provided. MrDocs can only identify whether a symbol is ultimated documented after extracting information from all translation units. For this reason, when this option is set to false, it's still recommendable to provide file and symbol filters so that only the desired symbols are traversed and stored by MrDocs.

Type: boolean
Default value: true

Example

Input

/// A documented function.
void documented();

// An undocumented function. The default `extract-all: true`
// would pull it in even without a doc comment; setting
// `extract-all: false` hides it.
void undocumented();

mrdocs.yml

extract-all: false

Preview·extract-all

documented

A documented function.

Synopsis

Declared in <extract‐all.cpp>

void
documented();

extract-private

Extraction policy for private class members

Determine whether private class members should be extracted

Type: boolean
Default value: false

Example

Input

/// A class with both public and private members.
class widget
{
public:
    /// Display the widget.
    void show();

private:
    /// An internal counter.
    int count;
};

mrdocs.yml

extract-private: true

Preview·extract-private

A class with both public and private members.

Synopsis

Declared in <extract‐private.cpp>

class widget;

Member Functions

Name

Description

show

Display the widget.

Private Data Members

Name

Description

count

An internal counter.

Display the widget.

Synopsis

Declared in <extract‐private.cpp>

void
show();

An internal counter.

Synopsis

Declared in <extract‐private.cpp>

int count;

extract-private-virtual

Extraction policy for private virtual methods of a class

Determine whether private virtual methods of a class should be extracted

Type: boolean
Default value: false

Example

Input

/// A base class declaring a virtual `draw` method.
struct shape
{
    virtual ~shape() = default;
    /// Render the shape.
    virtual void draw() = 0;
};

/// A circle implementation.
class circle : public shape
{
public:
    /// Construct a circle of the given radius.
    explicit circle(double r);

private:
    /// The private override of `shape::draw`.
    void draw() override;

    double radius;
};

mrdocs.yml

extract-private-virtual: true

Preview·extract-private-virtual

circle

A circle implementation.

Synopsis

Declared in <extract‐private‐virtual.cpp>

class circle
    : public shape

Base Classes

Name

Description

shape

A base class declaring a virtual draw method.

Member Functions

Name

Description

circle [constructor]

Construct a circle of the given radius.

draw [virtual]

Render the shape.

Private Member Functions

Name

Description

draw [virtual]

The private override of shape::draw.

circle::circle

Construct a circle of the given radius.

Synopsis

Declared in <extract‐private‐virtual.cpp>

explicit
circle(double r);

Parameters

Name

Description

The value to construct from

circle::draw

The private override of shape::draw.

Synopsis

Declared in <extract‐private‐virtual.cpp>

virtual
void
draw() override;

shape

A base class declaring a virtual draw method.

Synopsis

Declared in <extract‐private‐virtual.cpp>

struct shape;

Member Functions

Name

Description

~shape [destructor] [virtual]

Destructor

draw [virtual]

Render the shape.

Derived Classes

Name

Description

circle

A circle implementation.

shape::~shape

Destructor

Synopsis

Declared in <extract‐private‐virtual.cpp>

constexpr
virtual
~shape() = default;

shape::draw

Render the shape.

Synopsis

Declared in <extract‐private‐virtual.cpp>

virtual
void
draw() = 0;

extract-private-bases

Extraction policy for private base classes

Determine whether private base classes should be extracted

Type: boolean
Default value: false

Example

Input

/// A logger mixin.
class logger
{
public:
    /// Write a message to the log.
    void log(const char* msg);
};

/// A widget that privately inherits from `logger`
/// for the logging implementation.
class widget : private logger
{
public:
    /// Render the widget and write a log entry.
    void show();
};

mrdocs.yml

extract-private-bases: true

Preview·extract-private-bases

logger

A logger mixin.

Synopsis

Declared in <extract‐private‐bases.cpp>

class logger;

Member Functions

Name

Description

log

Write a message to the log.

logger::log

Write a message to the log.

Synopsis

Declared in <extract‐private‐bases.cpp>

void
log(char const* msg);

A widget that privately inherits from logger for the logging implementation.

Synopsis

Declared in <extract‐private‐bases.cpp>

class widget
    : logger

Member Functions

Name

Description

show

Render the widget and write a log entry.

Synopsis

Declared in <extract‐private‐bases.cpp>

void
show();

extract-static

Extraction policy for static members of a file

Determine whether static members of a file should be extracted. This option does not refer to static members of a class.

Type: boolean
Default value: false

Example

Input

/// A regular free function.
void run();

/// A static helper with internal linkage. With
/// `extract-static: true` it appears in the docs.
static void warm_caches();

mrdocs.yml

extract-static: true

Preview·extract-static

run

A regular free function.

Synopsis

Declared in <extract‐static.cpp>

void
run();

warm_caches

A static helper with internal linkage. With extract‐static: true it appears in the docs.

Synopsis

Declared in <extract‐static.cpp>

static
void
warm_caches();

extract-local-classes

Extraction policy for records defined locally in source files

Determine whether records only defined locally in source files should be extracted.

Type: boolean
Default value: true

Example

Input

/// A function whose local helper class is part of
/// the public interface (its instances escape via
/// the returned visitor).
inline void each()
{
    /// Local visitor type.
    struct visitor
    {
        void operator()(int) {}
    };
    visitor{}(0);
}

mrdocs.yml

extract-local-classes: true

Preview·extract-local-classes

each

A function whose local helper class is part of the public interface (its instances escape via the returned visitor).

Synopsis

Declared in <extract‐local‐classes.cpp>

void
each();

extract-anonymous-namespaces

Extraction policy for anonymous namespaces

Determine whether symbols in anonymous namespaces should be extracted.

Type: boolean
Default value: true

Example

Input

namespace lib {

/// A regular public function.
void run();

namespace {
/// A helper with translation-unit-local linkage. With
/// `extract-anonymous-namespaces: true` it shows up
/// in the docs.
void warm_caches();
}

}

mrdocs.yml

extract-anonymous-namespaces: true

Preview·extract-anonymous-namespaces

lib::run

A regular public function.

Synopsis

Declared in <extract‐anonymous‐namespaces.cpp>

void
run();

lib::warm_caches

A helper with translation‐unit‐local linkage. With extract‐anonymous‐namespaces: true it shows up in the docs.

Synopsis

Declared in <extract‐anonymous‐namespaces.cpp>

void
warm_caches();

extract-empty-namespaces

Extraction policy for empty namespaces

Determine whether empty namespaces without documentation should be extracted.

Type: boolean
Default value: false

Example

Input

namespace lib {
/** Greet the user by name. */
void greet(char const* name);
}

// Empty namespace reserved for future expansion. With
// `extract-empty-namespaces: true` it still appears in the docs.
namespace lib::experimental {}

mrdocs.yml

extract-empty-namespaces: true
show-namespaces: true

Preview·extract-empty-namespaces

Global namespace

Namespaces

Name

lib

lib

Namespaces

Name

experimental

Functions

Name

Description

greet

Greet the user by name.

lib::experimental

lib::greet

Greet the user by name.

Synopsis

Declared in <extract‐empty‐namespaces.cpp>

void
greet(char const* name);

inherit-base-members

Determine how derived classes inherit base members

Determine how derived classes inherit members of base classes. When set to never, derived classes do not inherit members of base classes and only the relationship is stored. When set to reference, derived classes list members of base classes but references are still linked to the base class. When set to copy-dependencies, a reference is created by default and a copy is created when the base class is a dependency. When set to copy-all, a copy is created for each base symbol as if it was declared in the derived class. If the base class is a dependency, the extraction mode is copied from the new parent.

Type: enum
Default value: "copy-dependencies"
Allowed values: ["never", "reference", "copy-dependencies", "copy-all"]

Example

Input

/// A base shape with one operation.
struct shape
{
    /// Render the shape to the screen.
    void draw();
};

/// A circle. The `draw` from `shape` is inherited
/// and listed on the circle's page too.
struct circle : shape
{
    /// Construct a circle of the given radius.
    explicit circle(double r);
};

mrdocs.yml

inherit-base-members: copy-all

Preview·inherit-base-members

circle

A circle. The draw from shape is inherited and listed on the circle's page too.

Synopsis

Declared in <inherit‐base‐members.cpp>

struct circle
    : shape

Base Classes

Name

Description

shape

A base shape with one operation.

Member Functions

Name

Description

circle [constructor]

Construct a circle of the given radius.

draw

Render the shape to the screen.

circle::circle

Construct a circle of the given radius.

Synopsis

Declared in <inherit‐base‐members.cpp>

explicit
circle(double r);

Parameters

Name

Description

The value to construct from

circle::draw

Render the shape to the screen.

Synopsis

Declared in <inherit‐base‐members.cpp>

void
draw();

shape

A base shape with one operation.

Synopsis

Declared in <inherit‐base‐members.cpp>

struct shape;

Member Functions

Name

Description

draw

Render the shape to the screen.

Derived Classes

Name

Description

circle

A circle. The draw from shape is inherited and listed on the circle's page too.

shape::draw

Render the shape to the screen.

Synopsis

Declared in <inherit‐base‐members.cpp>

void
draw();

extract-implicit-specializations

Implicit template specializations used as base classes are extracted as dependencies

When set to true, MrDocs extracts implicit template specializations used as base classes as dependencies. This allows MrDocs to extract metadata that can later be used to determine the members of the derived class, as specified by the inherit-base-members option.

Type: boolean
Default value: true

Example

Input

/// A simple container template.
template <class T>
struct box
{
    /// The contained value.
    T value;
};

/// Uses `box<int>`. With
/// `extract-implicit-specializations: true`, that
/// implicit instantiation shows up in the docs as
/// `box<int>` alongside the primary template.
box<int> make_int_box(int v);

mrdocs.yml

extract-implicit-specializations: true

Preview·extract-implicit-specializations

box

A simple container template.

Synopsis

Declared in <extract‐implicit‐specializations.cpp>

template<class T>
struct box;

Data Members

Name

Description

value

The contained value.

Non-Member Functions

Name

Description

make_int_box

Uses box<int>. With extract‐implicit‐specializations: true, that implicit instantiation shows up in the docs as box<int> alongside the primary template.

box::value

The contained value.

Synopsis

Declared in <extract‐implicit‐specializations.cpp>

T value;

make_int_box

Uses box<int>. With extract‐implicit‐specializations: true, that implicit instantiation shows up in the docs as box<int> alongside the primary template.

Synopsis

Declared in <extract‐implicit‐specializations.cpp>

box<int>
make_int_box(int v);

Return Value

A simple container template.

extract-friends

Extraction policy for friend functions and classes

Determine whether friend functions and classes should be extracted. When set to true, MrDocs extracts friend functions and classes. When set to false, friend functions and classes are not extracted.

Type: boolean
Default value: true

Example

Input

/// A point in 2D.
struct point
{
    double x;
    double y;

    /// Equality comparison, declared as a friend so it
    /// participates in ADL.
    friend bool operator==(point a, point b);
};

mrdocs.yml

extract-friends: true

Preview·extract-friends

point

A point in 2D.

Synopsis

Declared in <extract‐friends.cpp>

struct point;

Data Members

Name

x

y

Friends

Name

Description

operator==

Equality comparison, declared as a friend so it participates in ADL.

point::x

Synopsis

Declared in <extract‐friends.cpp>

double x;

point::y

Synopsis

Declared in <extract‐friends.cpp>

double y;

operator==

Equality comparison, declared as a friend so it participates in ADL.

Synopsis

Declared in <extract‐friends.cpp>

bool
operator==(
    point a,
    point b);

Return Value

true if the objects are equal, false otherwise

Parameters

Name

Description

The left operand

The right operand

sort-members

Sort the members of a record

When set to true, sort the members of a record by the criterion determined in the sort-members-by option. When set to false, the members are included in the declaration order they are extracted.

Type: boolean
Default value: true

Example

Input

/// A class declared with members in deliberately
/// scattered order to show how sorting affects the
/// rendered member list.
struct gadget
{
    /// Refresh the gadget.
    void refresh();
    /// Construct a gadget.
    gadget();
    /// Hide the gadget.
    void hide();
    /// Show the gadget.
    void show();
    /// Destroy the gadget.
    ~gadget();
};

mrdocs.yml

sort-members: true

Preview·sort-members

gadget

A class declared with members in deliberately scattered order to show how sorting affects the rendered member list.

Synopsis

Declared in <sort‐members.cpp>

struct gadget;

Member Functions

Name

Description

gadget [constructor]

Construct a gadget.

~gadget [destructor]

Destroy the gadget.

hide

Hide the gadget.

refresh

Refresh the gadget.

show

Show the gadget.

gadget::gadget

Construct a gadget.

Synopsis

Declared in <sort‐members.cpp>

gadget();

gadget::~gadget

Destroy the gadget.

Synopsis

Declared in <sort‐members.cpp>

~gadget();

gadget::hide

Hide the gadget.

Synopsis

Declared in <sort‐members.cpp>

void
hide();

gadget::refresh

Refresh the gadget.

Synopsis

Declared in <sort‐members.cpp>

void
refresh();

gadget::show

Show the gadget.

Synopsis

Declared in <sort‐members.cpp>

void
show();

sort-members-by

Determine how members of a record are sorted

If sort-members is set to true, determine how members of a record are sorted. When set to name, members are sorted by name. When set to location, members are sorted by their primary location in the source code, considering the short name of the path and the location in the file.

Type: enum
Default value: "name"
Allowed values: ["name", "location"]

Example

Input

/// Order members alphabetically by name (instead of
/// the default declaration order).
struct widget
{
    /// Refresh the widget.
    void refresh();
    /// Hide the widget.
    void hide();
    /// Show the widget.
    void show();
};

mrdocs.yml

sort-members: true
sort-members-by: name

Preview·sort-members-by

Order members alphabetically by name (instead of the default declaration order).

Synopsis

Declared in <sort‐members‐by.cpp>

struct widget;

Member Functions

Name

Description

hide

Hide the widget.

refresh

Refresh the widget.

show

Show the widget.

Hide the widget.

Synopsis

Declared in <sort‐members‐by.cpp>

void
hide();

Refresh the widget.

Synopsis

Declared in <sort‐members‐by.cpp>

void
refresh();

Show the widget.

Synopsis

Declared in <sort‐members‐by.cpp>

void
show();

sort-namespace-members-by

Determine how members of a namespace are sorted

Although members of namespaces are always sorted, determine how members of a namespace are sorted. When set to name, members are sorted by name. When set to location, members are sorted by their primary location in the source code, considering the short name of the path and the location in the file.

Type: enum
Default value: "name"
Allowed values: ["name", "location"]

Example

Input

namespace lib {

/// Initialize the library.
void init();
/// Connect to a remote service.
void connect();
/// Shut everything down.
void shutdown();

}

mrdocs.yml

sort-members: true
sort-namespace-members-by: name

Preview·sort-namespace-members-by

lib::connect

Connect to a remote service.

Synopsis

Declared in <sort‐namespace‐members‐by.cpp>

void
connect();

lib::init

Initialize the library.

Synopsis

Declared in <sort‐namespace‐members‐by.cpp>

void
init();

lib::shutdown

Shut everything down.

Synopsis

Declared in <sort‐namespace‐members‐by.cpp>

void
shutdown();

sort-members-ctors-1st

Sort constructors first

When set to true, constructors are sorted first in the list of members of a record.

Type: boolean
Default value: true

Example

Input

/// A class whose constructors should head the
/// rendered member list regardless of where the
/// declarations sit in the source.
struct widget
{
    /// Hide the widget.
    void hide();
    /// Show the widget.
    void show();
    /// Construct an empty widget.
    widget();
    /// Construct a widget with the given label.
    widget(const char* label);
};

mrdocs.yml

sort-members: true
sort-members-ctors-1st: true

Preview·sort-members-ctors-1st

A class whose constructors should head the rendered member list regardless of where the declarations sit in the source.

Synopsis

Declared in <sort‐members‐ctors‐1st.cpp>

struct widget;

Member Functions

Name

Description

widget [constructor]

Constructors

hide

Hide the widget.

show

Show the widget.

Constructors

Synopses

Declared in <sort‐members‐ctors‐1st.cpp>

Construct an empty widget.

widget();

» more...

Construct a widget with the given label.

widget(char const* label);

» more...

Construct an empty widget.

Synopsis

Declared in <sort‐members‐ctors‐1st.cpp>

widget();

Construct a widget with the given label.

Synopsis

Declared in <sort‐members‐ctors‐1st.cpp>

widget(char const* label);

Parameters

Name

Description

label

The value to construct from

Hide the widget.

Synopsis

Declared in <sort‐members‐ctors‐1st.cpp>

void
hide();

Show the widget.

Synopsis

Declared in <sort‐members‐ctors‐1st.cpp>

void
show();

sort-members-dtors-1st

Sort destructors first

When set to true, destructors are sorted first in the list of members of a record.

Type: boolean
Default value: true

Example

Input

/// A class whose destructor sits at the top of the
/// member list.
struct widget
{
    /// Refresh the widget.
    void refresh();
    /// Construct an empty widget.
    widget();
    /// Destroy the widget and release resources.
    ~widget();
};

mrdocs.yml

sort-members: true
sort-members-dtors-1st: true

Preview·sort-members-dtors-1st

A class whose destructor sits at the top of the member list.

Synopsis

Declared in <sort‐members‐dtors‐1st.cpp>

struct widget;

Member Functions

Name

Description

widget [constructor]

Construct an empty widget.

~widget [destructor]

Destroy the widget and release resources.

refresh

Refresh the widget.

Construct an empty widget.

Synopsis

Declared in <sort‐members‐dtors‐1st.cpp>

widget();

Destroy the widget and release resources.

Synopsis

Declared in <sort‐members‐dtors‐1st.cpp>

~widget();

Refresh the widget.

Synopsis

Declared in <sort‐members‐dtors‐1st.cpp>

void
refresh();

sort-members-assignment-1st

Sort assignment operators first

When set to true, assignment operators are sorted first in the list of members of a record.

Type: boolean
Default value: true

Example

Input

/// A class whose assignment operators should be
/// grouped at the top of the member list.
struct widget
{
    /// Refresh the widget.
    void refresh();
    /// Copy-assign from another widget.
    widget& operator=(const widget&);
    /// Move-assign from another widget.
    widget& operator=(widget&&);
};

mrdocs.yml

sort-members: true
sort-members-assignment-1st: true

Preview·sort-members-assignment-1st

A class whose assignment operators should be grouped at the top of the member list.

Synopsis

Declared in <sort‐members‐assignment‐1st.cpp>

struct widget;

Member Functions

Name

Description

operator=

Assignment operators

refresh

Refresh the widget.

Assignment operators

Synopses

Declared in <sort‐members‐assignment‐1st.cpp>

Copy‐assign from another widget.

widget&
operator=(widget const& other);

» more...

Move‐assign from another widget.

widget&
operator=(widget&& other);

» more...

Copy‐assign from another widget.

Synopsis

Declared in <sort‐members‐assignment‐1st.cpp>

widget&
operator=(widget const& other);

Return Value

Reference to the current object

Parameters

Name

Description

other

The object to copy assign from

Move‐assign from another widget.

Synopsis

Declared in <sort‐members‐assignment‐1st.cpp>

widget&
operator=(widget&& other);

Return Value

Reference to the current object

Parameters

Name

Description

other

The object to move assign from

Refresh the widget.

Synopsis

Declared in <sort‐members‐assignment‐1st.cpp>

void
refresh();

sort-members-conversion-last

Sort conversion operators last

When set to true, conversion operators are sorted last in the list of members of a record or namespace.

Type: boolean
Default value: true

Example

Input

/// A class whose conversion operators are pushed to
/// the bottom of the member list.
struct value
{
    /// Read the underlying integer.
    int get() const;
    /// Convert to `bool` for use in conditions.
    explicit operator bool() const;
    /// Convert to `int` for arithmetic contexts.
    explicit operator int() const;
};

mrdocs.yml

sort-members: true
sort-members-conversion-last: true

Preview·sort-members-conversion-last

value

A class whose conversion operators are pushed to the bottom of the member list.

Synopsis

Declared in <sort‐members‐conversion‐last.cpp>

struct value;

Member Functions

Name

Description

get

Read the underlying integer.

operator bool

Convert to bool for use in conditions.

operator int

Convert to int for arithmetic contexts.

value::get

Read the underlying integer.

Synopsis

Declared in <sort‐members‐conversion‐last.cpp>

int
get() const;

value::operator bool

Convert to bool for use in conditions.

Synopsis

Declared in <sort‐members‐conversion‐last.cpp>

explicit
operator bool() const;

Return Value

The object converted to bool

value::operator int

Convert to int for arithmetic contexts.

Synopsis

Declared in <sort‐members‐conversion‐last.cpp>

explicit
operator int() const;

Return Value

The object converted to int

sort-members-relational-last

Sort relational operators last

When set to true, relational operators are sorted last in the list of members of a record or namespace.

Type: boolean
Default value: true

Example

Input

/// A class whose relational operators sit at the
/// bottom of the member list.
struct value
{
    /// Get the underlying integer.
    int get() const;
    /// Equality comparison.
    bool operator==(const value&) const;
    /// Less-than comparison.
    bool operator<(const value&) const;
};

mrdocs.yml

sort-members: true
sort-members-relational-last: true

Preview·sort-members-relational-last

value

A class whose relational operators sit at the bottom of the member list.

Synopsis

Declared in <sort‐members‐relational‐last.cpp>

struct value;

Member Functions

Name

Description

get

Get the underlying integer.

operator==

Equality comparison.

operator<

Less‐than comparison.

value::get

Get the underlying integer.

Synopsis

Declared in <sort‐members‐relational‐last.cpp>

int
get() const;

Return Value

the underlying integer.

value::operator==

Equality comparison.

Synopsis

Declared in <sort‐members‐relational‐last.cpp>

bool
operator==(value const& rhs) const;

Return Value

true if the objects are equal, false otherwise

Parameters

Name

Description

rhs

The right operand

value::operator<

Less‐than comparison.

Synopsis

Declared in <sort‐members‐relational‐last.cpp>

bool
operator<(value const& rhs) const;

Return Value

true if the left object is less than the right object, false otherwise

Parameters

Name

Description

rhs

The right operand

Semantic Constructs

C++ semantic constructs to extract

Semantic constructs are patterns not directly represented in the source code AST but can be inferred from the corpus, such as SFINAE.

Name Description Default

sfinae
(boolean) Detect and reduce SFINAE expressions true

overloads
(boolean) Detect and group function overloads true

sfinae

Detect and reduce SFINAE expressions

When set to true, MrDocs detects SFINAE expressions in the source code and extracts them as part of the documentation. Expressions such as std::enable_if<...> are detected, removed, and documented as a requirement. MrDocs uses an algorithm that extracts SFINAE infomation from types by identifying inspecting the primary template and specializations to detect the result type and the controlling expressions in a specialization.

Type: boolean
Default value: true

Example

Input

#include <type_traits>

/** Multiply by two, but only for integers.

    The `std::enable_if_t` SFINAE on the return type
    restricts the overload to integral arguments.
*/
template <class T>
std::enable_if_t<std::is_integral_v<T>, T>
twice(T x);

mrdocs.yml

sfinae: true

Preview·sfinae

twice

Multiply by two, but only for integers.

Synopsis

Declared in <sfinae.cpp>

template<class T>
T
twice(T x)
requires std::is_integral_v<T>;

Description

The std::enable_if_t SFINAE on the return type restricts the overload to integral arguments.

overloads

Detect and group function overloads

When set to true, MrDocs detects function overloads and groups them as a single symbol type. The documentation for this new symbol comes from the union of non-ambiguous metadata from the functions.

Type: boolean
Default value: true

Example

Input

/** Compute the absolute value of an integer.

    One overload takes `int` and the other `long`,
    so the documentation page lists both side by side.
*/
int abs(int x);

/// @copydoc abs(int)
long abs(long x);

mrdocs.yml

overloads: true

Preview·overloads

abs

Compute the absolute value of an integer.

Synopses

Declared in <overloads.cpp>

Compute the absolute value of an integer.

int
abs(int x);

» more...

Compute the absolute value of an integer.

long
abs(long x);

» more...

abs

Compute the absolute value of an integer.

Synopsis

Declared in <overloads.cpp>

int
abs(int x);

Description

One overload takes int and the other long, so the documentation page lists both side by side.

abs

Compute the absolute value of an integer.

Synopsis

Declared in <overloads.cpp>

long
abs(long x);

Description

One overload takes int and the other long, so the documentation page lists both side by side.

Generators

Generators to create the documentation and their options

Name Description Default

generator
(enum) Generator used to create the documentation "adoc"

multipage
(boolean) Generate a multipage documentation true

base-url
(string) Base URL for links to source code

addons
(path) Path to the Addons directory "<mrdocs-root>/share/mrdocs/addons"

tagfile
(file path) Path for the tagfile "<output-dir>/reference.tag.xml"

legible-names
(boolean) Use legible names true

embedded
(boolean) Output an embeddable document false

show-namespaces
(boolean) Show namespace pages in the documentation true

show-enum-constants
(boolean) Show enum constant pages in the documentation false

global-namespace-index
(boolean) Use the global namespace page as an index for all symbols true

generator

Generator used to create the documentation

The generator is responsible for creating the documentation from the extracted symbols. The generator uses the extracted symbols and the templates to create the documentation. The generator can create different types of documentation such as HTML, XML, and AsciiDoc.

Type: enum
Default value: "adoc"
Allowed values: ["adoc", "html", "xml"]

Example

Input

/// Multiply two integers.
int multiply(int a, int b);

mrdocs.yml

generator: adoc

Preview·generator

multiply

Multiply two integers.

Synopsis

Declared in <generator.cpp>

int
multiply(
    int a,
    int b);

multipage

Generate a multipage documentation

Generates a multipage documentation. The output directory must be a directory. This option acts as a hint to the generator to create a multipage documentation. Whether the hint is followed or not depends on the generator.

Type: boolean
Default value: true

Example

mrdocs.yml

multipage: true

base-url

Base URL for links to source code

Base URL for links to source code. The base URL is used to create links to the source code in the documentation. The base URL is combined with the path to the source file to create the link.

Type: string
Default value:

Example

Input

/** Compute the area of a circle of radius `r`. */
double area(double r);

mrdocs.yml

base-url: https://github.com/example/geom/blob/main/

Preview·base-url

area

Compute the area of a circle of radius r.

Synopsis

Declared in <base‐url.cpp>

double
area(double r);

addons

Path to the Addons directory

Path to the Addons directory. The Addons directory contains the template files used by generators to create the documentation. When a custom Addons directory is not specified, the default templates are used. The default templates are located at the share/mrdocs/addons directory of the MrDocs installation. Users can create custom templates by copying the default templates to a custom directory and specifying the custom directory using this option.

Type: path
Default value: "<mrdocs-root>/share/mrdocs/addons"

tagfile

Path for the tagfile

Specifies the full path (filename) where the generated tagfile should be saved. If left empty, no tagfile will be generated.

Type: file path
Default value: "<output-dir>/reference.tag.xml"

legible-names

Use legible names

Use legible names for ids in the documentation. When set to true, MrDocs uses legible names for symbols in the documentation. These are symbols that are legible but still safe for URLs. When the option is set to false, MrDocs uses a hash of the symbol ID.

Type: boolean
Default value: true

Example

mrdocs.yml

multipage: true
legible-names: true

embedded

Output an embeddable document

Output an embeddable document, which excludes the header, the footer, and everything outside the body of the document. This option is useful for producing documents that can be inserted into an external template.

Type: boolean
Default value: false

show-namespaces

Show namespace pages in the documentation

When set to true, MrDocs creates a page for each namespace in the documentation.

Type: boolean
Default value: true

Example

Input

/// A tiny example library.
///
/// Holds the entry points the documentation calls out
/// and a few helpers consumed by the rest of the project.
namespace lib {

/// A function inside `lib`. With `show-namespaces: true`
/// the namespace appears in the rendered output as a
/// dedicated entry; with `false` the namespace is
/// flattened into the parent.
void run();

}

mrdocs.yml

show-namespaces: true

Preview·show-namespaces

Global namespace

Namespaces

Name

Description

lib

A tiny example library.

lib

A tiny example library.

Description

Holds the entry points the documentation calls out and a few helpers consumed by the rest of the project.

Functions

Name

Description

run

A function inside lib. With show‐namespaces: true the namespace appears in the rendered output as a dedicated entry; with false the namespace is flattened into the parent.

lib::run

A function inside lib. With show‐namespaces: true the namespace appears in the rendered output as a dedicated entry; with false the namespace is flattened into the parent.

Synopsis

Declared in <show‐namespaces.cpp>

void
run();

show-enum-constants

Show enum constant pages in the documentation

When set to true, MrDocs creates a page for each enum constant in the documentation.

Type: boolean
Default value: false

Example

Input

/// Status of a network operation.
enum class status
{
    /// Operation succeeded.
    ok,
    /// Operation timed out.
    timeout,
    /// Operation failed for another reason.
    error,
};

mrdocs.yml

show-enum-constants: true

Preview·show-enum-constants

status

Status of a network operation.

Synopsis

Declared in <show‐enum‐constants.cpp>

enum class status : int;

Members

Name

Description

ok

Operation succeeded.

timeout

Operation timed out.

error

Operation failed for another reason.

status::ok

Operation succeeded.

Synopsis

Declared in <show‐enum‐constants.cpp>

ok

status::timeout

Operation timed out.

Synopsis

Declared in <show‐enum‐constants.cpp>

timeout

status::error

Operation failed for another reason.

Synopsis

Declared in <show‐enum‐constants.cpp>

error

global-namespace-index

Use the global namespace page as an index for all symbols

When set to true, the page for the global namespace will recursively list all symbols in the documentation, not just those in the global namespace. This makes the global namespace page act as an index for the entire project.

Type: boolean
Default value: true

Build options

Options for building the source code

When MrDocs is responsible to running the build scripts and generating the compilation database, these options are used to build the source code.

Name Description Default

cmake
(string) CMake arguments when generating the compilation database from CMakeLists.txt

defines
(list of strings) Additional defines passed to the compiler []

use-system-stdlib
(boolean) Use the system C++ standard library false

stdlib-includes
(list of paths) C++ Standard Library include paths ["<mrdocs-root>/share/mrdocs/headers/libcxx", "<mrdocs-root>/share/mrdocs/headers/clang"]

use-system-libc
(boolean) Use the system C standard library false

libc-includes
(list of paths) Standard Library include paths ["<mrdocs-root>/share/mrdocs/headers/libc-stubs"]

system-includes
(list of paths) System include paths []

includes
(list of paths) Include paths []

cmake

CMake arguments when generating the compilation database from CMakeLists.txt

When the compilation-database option is a CMakeLists.txt file, these arguments are passed to the cmake command to generate the compilation_database.json.

Type: string
Default value:

defines

Additional defines passed to the compiler

Additional defines passed to the compiler when building the source code. These defines are added to the compilation database regardless of the strategy to generate it.

Type: list of strings
Default value: []

use-system-stdlib

Use the system C++ standard library

To achieve reproducible results, MrDocs bundles the LibC++ headers. To use the C++ standard library available in the system instead, set this option to true.

Type: boolean
Default value: false

stdlib-includes

C++ Standard Library include paths

When use-system-stdlib is disabled, the C++ standard library headers are available in these paths.

Type: list of paths
Default value: ["<mrdocs-root>/share/mrdocs/headers/libcxx", "<mrdocs-root>/share/mrdocs/headers/clang"]

use-system-libc

Use the system C standard library

To achieve reproducible results, MrDocs bundles the LibC headers with its definitions. To use the C standard library available in the system instead, set this option to true.

Type: boolean
Default value: false

libc-includes

Standard Library include paths

When use-system-libc is disabled, the C standard library headers are available in these paths.

Type: list of paths
Default value: ["<mrdocs-root>/share/mrdocs/headers/libc-stubs"]

system-includes

System include paths

System include paths. These paths are used to add directories to the system include search path. The system include search path is used to search for system headers. The system headers are headers that are provided by the system and are not part of the project. The system headers are used to provide the standard library headers and other system headers. The system headers are not part of the project and are not checked for warnings and errors.

Type: list of paths
Default value: []

includes

Include paths

Include paths. These paths are used to add directories to the include search path. The include search path is used to search for headers. The headers are used to provide declarations and definitions of symbols. The headers are part of the project and are checked for warnings and errors.

Type: list of paths
Default value: []

Warnings

Warnings and progress messages

Name Description Default

verbose
(boolean) Verbose output false

report
(unsigned integer)
(Deprecated) The minimum reporting level -1

log-level
(enum) The minimum reporting level "info"

warnings
(boolean) Enable warning messages true

warn-if-undocumented
(boolean) Warn if symbols are not documented true

warn-if-doc-error
(boolean) Warn if documentation has errors true

warn-no-paramdoc
(boolean) Warn if parameters are not documented true

warn-unnamed-param
(boolean) Warn if documented functions have unnamed parameters false

warn-if-undoc-enum-val
(boolean) Warn if enum values are not documented true

warn-broken-ref
(boolean) Warn if a documentation reference is broken true

warn-as-error
(boolean) Treat warnings as errors false

verbose

Verbose output

Verbose output. When set to true, MrDocs outputs additional information during the generation of the documentation.

Type: boolean
Default value: false

report

The minimum reporting level

The reporting level determines the amount of information displayed during the generation of the documentation. The value -1 delegates the decision to the log-level option.

Deprecated: Use log-level instead
Type: unsigned integer
Default value: -1
Minimum value: -1
Maximum value: 5

log-level

The minimum reporting level

The reporting level determines the amount of information displayed during the generation of the documentation.

Type: enum
Default value: "info"
Allowed values: ["trace", "debug", "info", "warn", "error", "fatal"]

warnings

Enable warning messages

When set to true, MrDocs outputs warning messages during the generation of the documentation. It is usually recommended to enable warnings while writing the documentation.

Type: boolean
Default value: true

warn-if-undocumented

Warn if symbols are not documented

When set to true, MrDocs outputs a warning message if a symbol that passes all filters is not documented.

Type: boolean
Default value: true

warn-if-doc-error

Warn if documentation has errors

When set to true, MrDocs outputs a warning message if the documentation of a symbol has errors such as duplicate parameters and parameters that don't exist.

Type: boolean
Default value: true

warn-no-paramdoc

Warn if parameters are not documented

When set to true, MrDocs outputs a warning message if a named function parameter is not documented.

Type: boolean
Default value: true

warn-unnamed-param

Warn if documented functions have unnamed parameters

When set to true, MrDocs outputs a warning message if a documented function has a parameter that is not named.

Type: boolean
Default value: false

warn-if-undoc-enum-val

Warn if enum values are not documented

When set to true, MrDocs outputs a warning message if an enum value is not documented.

Type: boolean
Default value: true

warn-broken-ref

Warn if a documentation reference is broken

When set to true, MrDocs outputs a warning message if a reference in the documentation is broken.

Type: boolean
Default value: true

warn-as-error

Treat warnings as errors

When set to true, MrDocs treats warnings as errors and stops the generation of the documentation.

Type: boolean
Default value: false

Miscellaneous

Miscellaneous options

Name Description Default

concurrency
(unsigned integer)
(Command line only) Number of threads to use 0

ignore-map-errors
(boolean) Continue if files are not mapped correctly false

ignore-failures
(boolean) Whether AST visitation failures should not stop the program false

concurrency

Number of threads to use

The desired level of concurrency: 0 for hardware-suggested.

Type: unsigned integer
Command line only
Default value: 0
Minimum value: 0

ignore-map-errors

Continue if files are not mapped correctly

When set to true, MrDocs continues to generate the documentation even if some files are not mapped correctly. Files are not mapped correctly when the source file is not found or the compilation database does not contain the compiler flags for the source file.

Type: boolean
Default value: false

ignore-failures

Whether AST visitation failures should not stop the program

When set to true, MrDocs continues to generate the documentation even if there are AST visitation failures. AST visitation failures occur when the source code contains constructs that are not supported by MrDocs.

Type: boolean
Default value: false